Low carbohydrate, high fat diet impairs exercise economy and negates the performance benefit from intensified training in elite race walkers
Participants
n=29 elite (world-class) race walkers in three isoenergetic diet groups during 3 weeks of intensified training: high-CHO availability (HCHO, 8.6 g/kg/day CHO, 2.1 protein, 1.2 fat; n=9); the same macronutrients periodised within/between days (PCHO, n=10); ketogenic LCHF (<50 g/day CHO, 78% energy as fat, 2.1 g/kg/day protein; n=10).
Protocol
Controlled dietary intervention during 3 weeks of intensified training in elite race walkers, with three isoenergetic diets (from the abstract)
HCHO: high carbohydrate availability before, during and after training (8.6 g/kg/day CHO, 2.1 protein, 1.2 fat; n=9); PCHO: identical macronutrient intake periodised within or between days to alternate low and high CHO availability (n=10); LCHF: ketogenic low-CHO, high-fat diet (78% energy as fat, 2.1 g/kg/day protein; n=10)
Measures: VO2peak during race walking, whole-body fat oxidation during 2 h of walking at ~80% VO2peak, O2 cost at a speed approximating 20 km race pace, and 10 km race-walk time
Key Findings
VO₂peak rose in all three groups after the 3-week block (P < 0.001; 90% CI 2.55-5.20%) — the diet did not change that.
LCHF markedly increased whole-body fat oxidation, peaking at 1.57 ± 0.32 g/min during 2 h of walking at ~80% VO₂peak.
LCHF impaired economy at race-relevant speed: O₂ uptake (as % of the new VO₂peak) at ~20 km race pace FELL in HCHO (90% CI -7.05 to -2.55%) and PCHO (-5.18 to -0.86%) but stayed at pre-intervention level in LCHF — i.e. the fat-adapted athletes did not gain the economy improvement the others did.
10 km race-walk time improved 6.6% (90% CI 4.1-9.1%) in HCHO and 5.3% (3.4-7.2%) in PCHO, with no improvement in LCHF (-1.6%; -8.5 to 5.3%).
Authors' conclusion: despite a significant VO₂peak gain, adaptation to LCHF negated the performance benefit of the training block, "in part due to reduced exercise economy".
Mechanism
Inference, not tested in this paper: oxidising fat yields less ATP per litre of O₂ than oxidising carbohydrate, so a fat-dominant substrate mix raises the O₂ cost of the same speed or power. The abstract reports the economy result, not the P/O measurement.
Practical Application
Burke et al. 2017 (n=29 world-class race walkers, 3 weeks): the ketogenic group doubled down on fat oxidation (peak 1.57 g/min) but lost the economy improvement the high-CHO and periodised groups gained, and its 10 km time did not improve (-1.6%) while HCHO improved 6.6% and PCHO 5.3%. Endurance athletes racing at high intensity should not adopt chronic LCHF in a training block they want to convert into performance.
Periodised CHO (train-low sessions inside a high-CHO framework) was NOT impaired — it improved economy and 10 km time like the chronic high-CHO diet. Selective low-CHO sessions remain a legitimate tool; a chronic ketogenic diet is a different thing.
Transfer: the study is race walking; the economy mechanism is not sport-specific (Inference, not tested in cyclists here).
Limitations
Specific to race walkers, not cyclists — but the underlying P/O / substrate-efficiency mechanism is sport-agnostic
3-week intervention — does not address whether much longer keto adaptation would change the economy result
n=29 in three groups (9/10/10); allocation method not given in the abstract; elite cohort, so transfer to recreational athletes pursuing weight-loss-driven LCHF is uncertain
How this entry was checked and changed
Why this level
regraded 2026-09-27 from A: parallel controlled trial, n=29 (9/10/10)
Content correction 2026-09-27
Level A -> B+: parallel controlled trial, n=29; "randomised" removed (allocation not stated in the abstract). Sport line cycling -> race walking (the study's discipline); n≈21 -> 29 (9/10/10).
"+5-8% O₂ cost" replaced by the abstract's result: O₂ cost fell in HCHO and PCHO (90% CIs given) and was unchanged in LCHF. Added the 10 km results (6.6% / 5.3% / -1.6%), VO₂peak rise in all groups, peak fat oxidation 1.57 g/min.
Deleted the garbled "5.05 kcal/L … ~6.3% more efficient" mechanism text and the P/O claim as a finding (now inference); "60-120 g/h" fuelling relabelled as the base's own practice. REF-461 byline Roberts -> Reynolds.
2026-09-28: added Protocol from the abstract (Europe PMC/PubMed PMID 28012184).